| OLD | NEW |
| 1 // Copyright (c) 2011, the Dart project authors. Please see the AUTHORS file | 1 // Copyright (c) 2011, the Dart project authors. Please see the AUTHORS file |
| 2 // for details. All rights reserved. Use of this source code is governed by a | 2 // for details. All rights reserved. Use of this source code is governed by a |
| 3 // BSD-style license that can be found in the LICENSE file. | 3 // BSD-style license that can be found in the LICENSE file. |
| 4 | 4 |
| 5 interface Scanner { | 5 interface Scanner { |
| 6 Token tokenize(); | 6 Token tokenize(); |
| 7 } | 7 } |
| 8 | 8 |
| 9 /** | 9 /** |
| 10 * Common base class for a Dart scanner. | 10 * Common base class for a Dart scanner. |
| (...skipping 611 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 622 return next; | 622 return next; |
| 623 } | 623 } |
| 624 } | 624 } |
| 625 if (raw) { | 625 if (raw) { |
| 626 return tokenizeSingleLineRawString(next, q, start); | 626 return tokenizeSingleLineRawString(next, q, start); |
| 627 } else { | 627 } else { |
| 628 return tokenizeSingleLineString(next, q, start); | 628 return tokenizeSingleLineString(next, q, start); |
| 629 } | 629 } |
| 630 } | 630 } |
| 631 | 631 |
| 632 static bool isHexDigit(int character) { |
| 633 if ($0 <= character && character <= $9) return true; |
| 634 character |= 0x20; |
| 635 return ($a <= character && character <= $f); |
| 636 } |
| 637 |
| 632 int tokenizeSingleLineString(int next, int q1, int start) { | 638 int tokenizeSingleLineString(int next, int q1, int start) { |
| 633 while (next !== $EOF) { | 639 loop: while (next !== $EOF) { |
| 634 if (next === q1) { | 640 if (next === q1) { |
| 635 appendByteStringToken(STRING_INFO, utf8String(start, 0)); | 641 appendByteStringToken(STRING_INFO, utf8String(start, 0)); |
| 636 return advance(); | 642 return advance(); |
| 637 } else if (next === $BACKSLASH) { | 643 } else if (next === $BACKSLASH) { |
| 638 next = advance(); | 644 next = advance(); |
| 639 if (next === $EOF) { | 645 if (next === $EOF || next === $LF || next === $CR) { |
| 640 throw new MalformedInputException(charOffset); | 646 throw new MalformedInputException(charOffset); |
| 641 } | 647 } |
| 648 switch (next) { |
| 649 case $u: |
| 650 next = advance(); |
| 651 bool isBracketed = false; |
| 652 int maxLength = 4; |
| 653 if (next === $OPEN_CURLY_BRACKET) { |
| 654 isBracketed = true; |
| 655 maxLength = 7; |
| 656 next = advance(); |
| 657 } |
| 658 int startOffset = charOffset; |
| 659 int unicodePoint = 0; |
| 660 int length = 0; |
| 661 while (length < maxLength && isHexDigit(next)) { |
| 662 int digit = next <= $9 ? next - 0x30 : (next | 0x20) - ($a - 10); |
| 663 unicodePoint = unicodePoint * 16 + digit; |
| 664 length += 1; |
| 665 next = advance(); |
| 666 } |
| 667 if (isBracketed) { |
| 668 if (length === 0) { |
| 669 throw new MalformedInputException(startOffset); |
| 670 } |
| 671 if (next !== $CLOSE_CURLY_BRACKET) { |
| 672 throw new MalformedInputException(charOffset); |
| 673 } |
| 674 next = advance(); |
| 675 } else if (length != 4) { |
| 676 throw new MalformedInputException(startOffset + length); |
| 677 } |
| 678 if ((0xD800 <= unicodePoint && unicodePoint <= 0xDFFF) || |
| 679 unicodePoint > 0x10ffff) { |
| 680 // Don't allow invalid code points in strings. |
| 681 throw new MalformedInputException(startOffset); |
| 682 } |
| 683 // TODO(lrn): This should be just "continue", but the VM is bugged. |
| 684 continue loop; |
| 685 case $x: |
| 686 for (int i = 0; i < 2; i++) { |
| 687 next = advance(); |
| 688 if (!isHexDigit(next)) { |
| 689 throw new MalformedInputException(charOffset); |
| 690 } |
| 691 } |
| 692 break; |
| 693 default: |
| 694 // All other escapes are trivially valid. |
| 695 } |
| 642 } else if (next === $$) { | 696 } else if (next === $$) { |
| 643 beginToken(); | 697 beginToken(); |
| 644 next = advance(); | 698 next = advance(); |
| 645 if (next === $OPEN_CURLY_BRACKET) { | 699 if (next === $OPEN_CURLY_BRACKET) { |
| 646 next = tokenizeInterpolatedExpression(next, start); | 700 next = tokenizeInterpolatedExpression(next, start); |
| 647 } else { | 701 } else { |
| 648 next = tokenizeInterpolatedIdentifier(next, start); | 702 next = tokenizeInterpolatedIdentifier(next, start); |
| 649 } | 703 } |
| 650 start = byteOffset; | 704 start = byteOffset; |
| 651 continue; | 705 continue; |
| (...skipping 57 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 709 } | 763 } |
| 710 return next; | 764 return next; |
| 711 } | 765 } |
| 712 } | 766 } |
| 713 | 767 |
| 714 class MalformedInputException { | 768 class MalformedInputException { |
| 715 final message; | 769 final message; |
| 716 MalformedInputException(this.message); | 770 MalformedInputException(this.message); |
| 717 toString() => message.toString(); | 771 toString() => message.toString(); |
| 718 } | 772 } |
| OLD | NEW |